C-LARA

An AI collaborates with humans to build a language learning app.

Weekly summary, Apr 18-24 2024

This week, I have been tidying things up. I carried on installing old LARA content on the UniSA server and fixed several annoying issues in the platform.

Legacy LARA content

We now have nearly all of the old LARA content installed on the UniSA server. I found that in some cases material had not been checked in correctly, but I was able to bring up my old laptop and retrieve things from it when necessary.

Fixing small problems

I have fixed a number of small but annoying problems that had been hanging around for too long:

  • In Simple C-LARA, there is now an extra step that produces the segmented text before proceeding to creation of multimedia text. Since the segmented text often needs editing, this saves considerable time.
  • Gloss popups should now work better. After some experimentation, I found that making them appear below the word rather than above and increasing the size of the margins improves things considerably.
  • Search now interacts correctly with pagination.
  • I have added functionality to cache annotations, so if a small change is made it is not necessary to redo all the annotation work.

Next Zoom call

The next call will be at:

Thu Apr 25 2024, 18:00 Adelaide (= 08.30 Iceland = 09.30 Ireland/Faroe Islands = 10.30 Europe = 11.30 Israel = 12.00 Iran = 16.30 China = 18:30 Melbourne = 19.30 New Caledonia)

The distinguished researcher Francis Bond says he is interested in finding out more about the project and may join us.

,
,
mannyrayner's avatar

Posted by:

mannyrayner

2 responses to “Weekly summary, Apr 18-24 2024”

  1. cathyc Avatar
    cathyc

    There is a lot of ‘I’ in this and not much of the ‘we’ your readers are more used to. Is this because Chat didn’t have much to do with the decisions or is it a mere human oversight?

    Like

    Reply
    1. mannyrayner Avatar
      mannyrayner

      In fact, this was for once mostly my work! I did all the restoration of the old LARA texts, which often depended on knowing various details about the projects, and the Simple C-LARA enhancement – as noted in the report, Chat finds Simple C-LARA the hardest part of the code to understand. The caching and the gloss popups were the result of discussions with the AI, but I ended up writing the code.

      However, Chat found a very nice, elegant solution to the problem of making search and pagination interact correctly. It used a trick I didn’t know and would probably not have found on my own.

      Liked by 1 person

      Reply

Leave a comment

Categories